Why Krugman Is Wrong About AI

Dear Reader,

“ChatGPT and whatever follows are probably an economic story for the 2030s, not for the next few years.” 

That’s a line from Paul Krugman’s op-ed in The New York Times.

In 2008, Krugman won a Nobel Prize in Economics. But you’d be hard-pressed to know why based on his past calls.

In 1998, he said, “By 2005 or so, it will become clear that the internet’s impact on the economy has been no greater than the fax machine’s.”

Ironically, this was a prediction he made in an article titled “Why most economists’ predictions are wrong.”

Other economists aside, he couldn’t have been more wrong himself.

By 2005, Amazon, Alphabet, PayPal, and Microsoft had a combined market cap of over $400 billion. Over 68% of the U.S. populace was online, according to Pew Research. 

Krugman wasn’t just wrong about tech, either.

In 2002, he said that American prosperity had reached its peak. Since 2002, America’s GDP has more than doubled from $10.9 trillion to $23.3 trillion.

In 2011, Krugman dismissed debt concerns in Europe. Even going as far as to say that Spain wouldn’t experience any major crisis. Within a year, European banks bailed out Spain with over $150 billion in aid.

If his track record is anything to go by, it’s a great sign for AI.

But let’s entertain Krugman’s arguments to see why the skeptics are so confident it’s all hype.

AI Is Like… the Steam Engine?

Krugman’s most recent article panning AI looks back 100 years at the slow transition from steam engines to electricity.

Here’s what he said.

To take full advantage of electrification, manufacturers had to rethink the design of factories. Pre-electric factories were multistory buildings with cramped working spaces, because that was necessary to make efficient use of a steam engine in the basement driving the machines through a system of shafts, gears, and pulleys.

It took time to realize that having each machine driven by its own motor made it possible to have sprawling one-story factories with wide aisles allowing easy movement of materials, not to mention assembly lines. As a result, the big productivity gains from electrification didn’t materialize until after World War I.

Let’s put aside the baffling comparison to an adoption cycle for a technology that kicked off a century ago.

The crux of Krugman’s argument is that it’ll take a long time for companies to recognize that AI could make companies more efficient and productive. And even longer for them to adopt it.

A recent survey of 1,600 C-suite executives found that 75% were developing an AI strategy or already using AI. So, it’s not as if businesses are unaware of the potential for artificial intelligence.

And it’s not like AI tools won’t be available for easy adoption.

Microsoft is already building AI capabilities into its products.

• It uses ChatGPT to improve search results from Bing.

• Wholly owned GitHub uses Co-pilot to help developers write code faster and more effectively.

• Azure has several AI tools to help businesses make workflow even more efficient.

Salesforce is another software company that helps businesses manage their data and customer relations. It already uses Einstein, an AI platform. And it’s rolling out AI in all of its services over the next year. That includes Slack, Tableau, and its cloud platforms.

Adopting AI is very different from converting a factory from steam power to electricity. The latter would have meant years of rebuilding physical infrastructure at an incredible cost. 

But current cloud-based software will allow these updates to be integrated seamlessly. No new factories need to be built. It’s “plug and play.” Users won’t even have to go out of their way to use AI.

And Krugman makes the mistake of thinking that the AI adoption cycle just started. In reality, the AI adoption we’re seeing now is the result of long-term preparation.

Since 2019, Microsoft has invested $13 billion into ChatGPT’s creator, OpenAI.

Together, Microsoft and OpenAI have spent hundreds of millions on Nvidia’s AI chips.

Salesforce launched its first AI initiative, Einstein, in 2016 after years of development.

Others aren’t far behind. Meta is developing its own AI tech. Meta will have generative AI tools built into Facebook, Instagram, and WhatsApp in the coming months.

Krugman seems to mistakenly think that today’s AI wasn’t the product of billions of dollars and years of development.

When Krugman was questioned about his wrong call on the internet’s impact, he responded, “I was clearly trying to be provocative and got it wrong, which happens to all of us sometimes.”

I believe that Krugman is trying to be provocative again. And I know he’s just as wrong about AI as he was about the internet in 1998.

There will be plenty of other pundits in the weeks and months ahead that will dismiss AI just for the attention. Don’t fall for it.

Regards,

Colin Tedards
Editor, The Bleeding Edge